There is accumulating evidence that chronic injury and inflammation in the respiratory tract, such as that caused by cigarette smoking, predispose to lung cancer. We therefore propose to conduct an in depth pathway-based analysis of gene variants in the inflammatory pathway using test and validation sets of cases and controls. This proposal builds on a well-annotated specimen repository of lung cancer cases and controls enrolled in an ongoing risk factor study (CA55769, Spitz, PI). Cases are frequency-matched to controls on age, gender, ethnicity and smoking status recruited from a multi-specialty physician practice. Data collected include smoking history, dietary intake, cancer family history, specific occupational exposures (e.g., asbestos, dust), and previous medical history including chronic obstructive airway disease, asthma and hay fever. Genomic DNA and rich candidate genotype and phenotype data are available. Aim 1: To identify novel genetic variants influencing lung cancer risk in a test set of 1500 cases with non-small cell lung cancer and 1500 matched controls (all Caucasian), using the Illumina iSelect Infinium chip with 8.5 to 9K SNP's. Aim 2: In a replication set of an additional 1000 cases and 1000 controls, using a GoldenGate assay, we will evaluate the top 1500 SNPs identified from Aim1 as meeting the P<0.1 criterion, or selected by a rational prioritizing approach that incorporates published results, type of SNP, evolutionary biology, physico-chemical properties and haplotype tagging SNPs. Aim 3: To perform fine mapping in the flanking regions of 50 SNPs selected by the same approach as in Aim 2, combining prior information with in silico approaches for predicting functionality. For each of these 50 SNPs, we will select an average of 10 additional SNPs per gene region to regenotype in all 2500 cases and 2500 controls. Aim 4. To extend our epidemiologic risk prediction model by incorporating established epidemiologic risk factor and gene variant data. We will apply machine-learning tools to identify gene-environment and gene-gene interactions. Covariates will include prior emphysema, asthma, hay fever, dust and asbestos exposure, smoking characteristics, family history of cancer, and anti-inflammatory drug use. The International Lung Cancer Consortium will perform external validation in a proposal to be developed. Our approach to comprehensively evaluate variants in a candidate pathway in a large well- powered study will be applicable to a variety of other cancer sites where inflammation plays an important etiologic role, as well as in non-neoplastic diseases with a strong inflammatory component such as emphysema. The public health potential of a useful risk prediction modes for lung cancer is substantial.